The future point the author makes is that all skill has been taken out of most jobs, and almost all power to boot. In the future, no one will be able to handle exceptions not caught by the software.
Right. Systems are only getting more complex. What happens when a wildly complex AI decides that you should not be sold a ticket to this movie or train, and there are no attendants around because why would there be? They haven't been needed for years, "the system does not make mistakes". What if it's an AI that has lethal measures to protect an area (etc etc)?
This is already a thing in China and their social credit system. Lose enough points and you can be prevented from buying a train ticket or booking a flight [0]. You can probably imagine that rectifying errors in this system is... troublesome.
